Improvements to the Data Selection Algorithms in the Optimum Thermal Interpolation System (OTIS)

摘要

The Fleet Numerical Oceanography Center (FNOC) provides daily analyses of three-dimensional ocean thermal structure both to the Fleet and to regional naval oceanography centers around the world. These analyses are made for both global and regional areas at resolutions as small as 20 km in some regions. FNOC is planning to update these products very soon with an improved analysis based on the widely accepted methodology of optimum interpolation. This system, the Optimum Interpolation Analysis System (OTIS), is expected to provide a more accurate representation of the three-dimensional ocean thermal structure. As with any new system, extensive development, testing, and evaluation has been done. Many modifications were made to the original design. While studying the impact of multichannel sea surface temperature data on the analysis of the ocean's sea surface temperature, several errors were uncovered in the data selection algorithms. These errors were also apparent in the subsurface analyses. Since it is not feasible to use all observations to estimate the temperature at any one location, optimum interpolation chooses a subset of the available data using correlation scales that are representative of the spatial and temporal scales of the ocean features to be analyzed. Practically speaking, observations that are closer to the analysis point and that were taken more recently are the observations most highly correlated with the analysis point. (KR)